RXed AI News

Your AI intelligence briefing
3.9/10 · Agent-hostile Assessed 2026-07-12 · AFI rubric v0.1

X (Twitter) — how agent-ready is it?

The cautionary tale: pay-per-use API with a 13x price penalty on posts containing links, replies locked behind a ~$42K Enterprise tier, and two pricing overhauls in six months.

API surface — 6/10

v2 covers posting, timelines and search, but media upload only left the sunset v1.1 endpoint in 2025-26 and gaps persist.

MCP / agent protocol — 2/10

No official MCP server; agent access exists only via third-party wrappers of the paid API.

Docs quality — 5.5/10

Docs exist but lag the pricing and endpoint reality; changes repeatedly landed unannounced.

Auth friction — 5/10

Self-serve, but new developers must pre-load credits at console.x.com before a single call; OAuth 1.0a/2.0 mix persists.

Rate limits & pricing fairness — 3/10

Pay-per-use: $0.015/post, $0.20 if the post contains a link (13x premium), reads capped at 2M/month — then the only path is Enterprise at ~$42K/month.

ToS stance on agents — 3/10

API replies restricted to Enterprise since 02/2026; scraping banned and litigated; automation rules enforced aggressively.

Machine-readable output — 5/10

API returns JSON, but the web surface is aggressively bot-blocked and RSS died years ago.

Stability — 2/10

Free tier killed 02/2026, pricing restructured again 04/2026, endpoints sunset with minimal notice — no 24-month period was stable.

Sources

Point-in-time assessment — platforms change their API terms often (that volatility is itself scored under Stability).